janhq/Jan-v3-4B-base-instruct
Partial fine-tune vs Qwen/Qwen3-4B-Instruct-2507. Partial change: 89% of the weights changed (attention, mlp, norms), in 36 of 36 layers.
Tensors changed: Embeddings 0/1, Norms 60/145, MLP 106/108, Attention 144/144
10 derivatives checked
- janhq/Jan-v3.5-4B-gguf: Full fine-tune
- janhq/Jan-v3-4B-base-instruct-gguf: Quantization
- bartowski/janhq_Jan-v3-4B-base-instruct-GGUF: Same weights, converted
- aashish1904/Jan-v3-4B-base-instruct-GGUF: Quantization
- janhq/Jan-v3.5-4B: Partial fine-tune
- mradermacher/Jan-v3-4B-base-instruct-i1-GGUF: Quantization
- mlx-community/Jan-v3-4B-base-instruct-4bit: Quantization
- mradermacher/Jan-v3-4B-base-instruct-GGUF: Same weights, converted
- janhq/Jan-code-4b: Full fine-tune
- megabytes/Jan-v3-4B-base-instruct-heretic: Abliterated